#9472 NORM Not Tri: memtest is broken in q3a11

[Zarro Boogs per Child]

#9472: memtest is broken in q3a11
-------------------------------------------+--------------------------------
           Reporter:  wmb at firmworks.com    |       Owner:  wmb at firmworks.com
               Type:  defect               |      Status:  new              
           Priority:  normal               |   Milestone:  Not Triaged      
          Component:  ofw - open firmware  |     Version:  not specified    
         Resolution:                       |    Keywords:                   
        Next_action:  never set            |    Verified:  0                
Deployment_affected:                       |   Blockedby:                   
           Blocking:                       |  
-------------------------------------------+--------------------------------

Comment(by wmb at firmworks.com):

 Part of the problem is that sequencer register 1a had bit 0 cleared, so
 the palette accesses were going to the primary display engine, which is
 turned off.  Setting that bit to 1 eliminates the hang and lets memtest
 run with output to serial, but screen output doesn't work.  That is
 because we don't have screen scaling working for the LCD in text mode 3.

 One solution would be to go back to the memtest version that does emulated
 text mode on the graphics frame buffer.  That would fix memtest, but there
 is another requirement for real text mode - Windows booting.  So it seems
 that I will need to make text mode work at some point.

-- 
Ticket URL: <http://dev.laptop.org/ticket/9472#comment:2>
One Laptop Per Child <http://laptop.org/>
OLPC bug tracking system